nan Suroso

Inverse Technique for Analysis of Convective Heat Transfer over the Surface of Plant Culture Vessel

A practical method to determine the forced convective heat transfer coefficient over the plant culture vessel was developed. A finite element neural network inverse technique was used to determine parameter values of the Nusselt equation, which, in turn, is needed to calculate the forced convective heat transfer coefficient. The values of the forced convective heat transfer coefficient obtained by the proposed method were then used for the finite element calculation. Thereby, the temperature distribution inside a plant culture vessel was simulated. Agreement was obtained between the finite element results and experimentally measured temperature distribution inside the plant culture vessel.

Non-Destructive Analysis for Citrus and Lanzone Fruit Qualities Using ANN

Abstract The objective was to develop a non-destructive determination of physical and chemical properties of citrus and lanzone fruits as a screening method for fruit quality differentiation. The results show that although the highest correlation were shown in juice, the correlation in the whole-fruit samples were very low. Consequently, the NIRS method was judged to be inappropriate for a non-destructive determination on the chemical qualities of citrus. The tested method of visible light with the light dependent resistor system could not predict seed number and weight in citrus as two important seediness parameters, but might still be applicable to lanzone fruits.

Classification of Mango by Artificial Neural Network Based on Near Infrared Diffuse Reflectance

Abstract The objective of this study was to apply the artificial neural network for mango classification based on their near infrared reflectance and the organoleptic taste score. The results indicated that the use of both the stepwise method and the principal component analysis in providing input for the neural network was feasible. The results of neural network on mango classification reached an accuracy of 88.9% for the sweet taste group. 100.0% for the sweet sour. 83.3% for the sour, and 100.0% for the bland ones at 5 principal Components .

Model for Predicting and Classifying Durian Fruit Based on Maturity and Ripeness Using Neural Network

Abstract This study was aimed to develop the model to predict the maturity, ripeness and defects of durian based on its physical and chemical characteristics by using the neural network. The density and acoustic characteristics measurement was fed into the model as the inputs, which provided the levels of maturity and ripeness as the output. Data training were tested to models of neural network with various nodes in the hidden layer, i.e., 4, 6, 8, and 10 nodes. The results recommended the use of 6 nodes in the hidden layer that would provide the highest accuration of 100 % in classifying the durian based on its maturity and ripeness.

Neural Network Application on Non-Destructive Evaluation of Lanzone Using Visible Light

Abstract A neural network model using 4 nodes in the hidden layer and 6000 iterations was applied to predict the weight of lanzone seeds in the fruit, thus, classifying the seedless from the seed ones. Inputs fed to the model were the wholefruit weight, the diameter and the transmittance intensity ratio of visible light passing through the fruits. The output was the weight of lanzone seeds inside the pulp. The model successfully predict the weight of the lanzone seeds with a determinant coefficient of 0.86.